21ANTHROPOLOGY IS NOT ETHNOGRAPHY

Professor Tim Ingold FBA gave the2007 Radcliffe-Brown Lecture in SocialAnthropology. In these edited extractsfrom his lecture, he reveals the differingviews on what anthropology andethnography are, and recalls some ofthe heated past debates about thesedifferences.

THE OBJECTIVE of anthropology, I believe,

is to seek a generous, comparative but

nevertheless critical understanding of human

being and knowing in the one world we all

inhabit. The objective of ethnography is to

describe the lives of people other than

ourselves, with an accuracy and sensitivity

honed by detailed observation and prolonged

first-hand experience. My thesis is that

anthropology and ethnography are

endeavours of quite different kinds. This is

not to claim that the one is more important

than the other, or more honourable. Nor is it

to deny that they depend on one another in

significant ways. It is simply to assert that

they are not the same. Indeed this might

seem like a statement of the obvious, and so

it would be were it not for the fact that it has

become commonplace – at least over the last

quarter of a century – for writers in our

subject to treat the two as virtually

equivalent, exchanging anthropology for

ethnography more or less on a whim, as the

mood takes them, or even exploiting the

supposed synonymy as a stylistic device to

avoid verbal repetition. Many colleagues to

whom I have informally put the question

have told me that in their view there is little

if anything to distinguish anthropological

from ethnographic work. Most are convinced

that ethnography lies at the core of what

anthropology is all about. For them, to

suggest otherwise seems almost anachron-

istic. It is like going back to the bad old

days – the days, some might say, of Radcliffe-

Brown. For it was he who, in laying the

foundations for what was then the new

science of social anthropology, insisted on

the absolute distinction between ethno-

graphy and anthropology.

He did so in terms of a contrast, much

debated then but little heard of today,

between idiographic and nomothetic inquiry.

An idiographic inquiry, Radcliffe-Brown

explained, aims to document the particular

facts of past and present lives, whereas the

aim of nomothetic inquiry is to arrive

at general propositions or theoretical

statements. Ethnography, then, is specifically

a mode of idiographic inquiry, differing from

history and archaeology in that it is based on

the direct observation of living people rather

than on written records or material remains

attesting to the activities of people in the

past. Anthropology, to the contrary, is a field

of nomothetic science. As Radcliffe-Brown

declared in his introduction to Structure and

Function in Primitive Society – in a famous

sentence that, as an undergraduate beginning

my anthropological studies at Cambridge in

the late 1960s, I was expected to learn by

heart – ‘comparative sociology, of which

social anthropology is a branch, is … a

theoretical or nomothetic study of which the

aim is to provide acceptable generalisations.’

This distinction between anthropology and

ethnography was one that brooked no

compromise, and Radcliffe-Brown reasserted

it over and over again. Returning to the

theme in his Huxley Memorial Lecture for

1951 on ‘The comparative method in social

anthropology’, best known for its revision of

the theory of totemism, Radcliffe-Brown

insisted that ‘without systematic comparative

studies anthropology will become only

historiography and ethnography.’ And the

aim of comparison, he maintained, is to pass

from the particular to the general, from the

general to the more general, and ultimately

to the universal.

On the other side of the Atlantic, however,

a very different call was being put out

by Radcliffe-Brown’s contemporary, Alfred

Kroeber, for an anthropology that would

be fully historical. The historical approach –

in Kroeber’s words – ‘finds its intellectual

satisfaction in putting each preserved

phenomenon into a relation of ever widening

context within the phenomenal cosmos.’ In

1935 he had characterised this task, of

preservation through contextualisation, as

‘an endeavour at descriptive integration’. As

such, it is of an entirely different kind from

the task of theoretical integration that

Radcliffe-Brown had assigned to social

anthropology. For the latter, in order to

generalise, must first isolate every particular

from its context in order that it can then be

subsumed under context-independent

formulations. Kroeber’s disdain for Radcliffe-

Brown’s understanding of history, as nothing

but a chronological tabulation of such

isolated particulars awaiting the classificatory

and comparative attentions of the theorist,

bordered on contempt. ‘I do not know the

motivation for Radcliffe-Brown’s depreciation

of the historical approach,’ he remarked

caustically in an article published in 1946,

‘unless that, as the ardent apostle of a

genuine new science of society, he has

perhaps failed to concern himself enough

with history to learn its nature.’

In Britain, Kroeber’s understanding of what a

historical or ideographic anthropology would

look like fell on the sympathetic ears of E. E.

Evans-Pritchard. In his Marett Lecture of

1950, ‘Social anthropology: past and present’,

Evans-Pritchard reiterated, almost word for

word, what Kroeber had written fifteen years

Anthropology is not ethnography

A. R. Radcliffe-Brown, Fellow of the British Academy

British Academy Review, issue 11. © The British Academy 2008

previously about the relation between

anthropology and history. This is what he

said:

I agree with Professor Kroeber that the

fundamental characteristic of the historical

method is not chronological relation of

events but descriptive integration of them;

and this characteristic historiography

shares with social anthropology. What

social anthropologists have in fact chiefly

been doing is to write cross-sections of

history, integrative descriptive accounts of

primitive peoples at a moment in time

which are in other respects like the

accounts written by historians about

peoples over a period of time…

Returning to this theme over a decade later,

in a lecture on ‘Anthropology and history’

delivered at the University of Manchester,

Evans-Pritchard roundly condemned – as had

Kroeber – the blinkered view of those such as

Radcliffe-Brown for whom history was

nothing more than ‘a record of a succession

of unique events’ and social anthropology

nothing less than ‘a set of general

propositions’. In practice, Evans-Pritchard

claimed, social anthropologists do not

generalise from particulars any more that do

historians. Rather, ‘they see the general in the

particular.’ Or to put it another way, the

singular phenomenon opens up as you go

deeper into it, rather than being eclipsed

from above.

The problem is that once the task of

anthropology is defined as descriptive

integration rather than comparative

generalisation, the distinction between

ethnography and social anthropology, on

which Radcliffe-Brown had set such store,

simply vanishes. Beyond ethnography, there

is nothing left for anthropology to do. And

Radcliffe-Brown himself was more than aware

of this. In a 1951 review of Evans-Pritchard’s

book Social Anthropology, in which the author

had propounded the same ideas about

anthropology and history as those set out in

his Marett lecture, Radcliffe-Brown registered

his strong disagreement with ‘the implication

that social anthropology consists entirely or

even largely of … ethnographic studies of

particular societies. It is towards some such

position that Professor Evans-Pritchard and a

few others seem to be moving.’ And it was

indeed towards such a position that the

discipline moved over the ensuing decade, so

much so that in his Malinowski Lecture of

1959, ‘Rethinking Anthropology’, Edmund

Leach felt moved to complain about it. ‘Most

of my colleagues’, he grumbled, ‘are giving up

in the attempt to make comparative

generalizations; instead they have begun to

write impeccably detailed historical

ethnographies of particular peoples.’ But did

Leach, in regretting this tendency, stand up

for the nomothetic social anthropology of

Radcliffe-Brown? Far from it. Though all in

favour of generalisation, Leach launched an

all-out attack on Radcliffe-Brown for having

gone about it in the wrong way. And the

source of the error, he maintained, lay not in

generalisation per se, but in comparison.

There are two varieties of generalisation,

Leach argued. One, the sort of which he

disapproved, works by comparison and

classification. It assigns the forms or

structures it encounters into types and

subtypes, as a botanist or zoologist, for

example, assigns plant or animal specimens

to genera and species. Radcliffe-Brown liked

to imagine himself working this way. As

he wrote in a letter to Claude Lévi-Strauss,

social structures are as real as the structures of

living organisms, and may be collected and

compared in much the same way in order to

arrive at ‘a valid typological classification’.

The other kind of generalisation, of which

Leach approved, works by exploring a priori –

or as he put it, by ‘inspired guesswork’ – the

space of possibility opened up by the

combination of a limited set of variables. A

generalisation, then, would take the form not

of a typological specification that would

enable us to distinguish societies of one kind

from those of another, but of a statement of

the relationships between variables that may

operate in societies of any kind. This is the

approach, Leach claimed, not of the botanist

or zoologist, but of the engineer. Engineers

are not interested in the classification of

machines, or in the delineation of taxa. They

want to know how machines work. The task

of social anthropology, likewise, is to

understand and explain how societies work.

Of course, societies are not machines, as

Leach readily admits. But if you want to find

out how societies work, they may just as well

be compared to machines as to organisms.

‘The entities we call societies’, Leach wrote,

‘are not naturally existing species, neither are

they man-made mechanisms. But the

analogy of a mechanism has quite as much

relevance as the analogy of an organism.’

I beg to differ, and on this particular point

I want to rise to the defence of Radcliffe-

Brown who, I think, has been wilfully

misrepresented by his critics, including both

Leach and Evans-Pritchard. According to

Leach, Radcliffe-Brown’s resort to the organic

analogy was based on dogma rather than

choice. Not so. It was based on Radcliffe-

Brown’s commitment to a philosophy of

process. On this he was absolutely explicit.

22 ANTHROPOLOGY IS NOT ETHNOGRAPHY

E. E. Evans-Pritchard, Fellow of the British Academy

Edmund Leach, Fellow of the British Academy

23ANTHROPOLOGY IS NOT ETHNOGRAPHY

Societies are not entities analogous to

organisms, let alone to machines. In reality,

indeed, there are no such entities. ‘My own

view’, Radcliffe-Brown asserted, ‘is that the

concrete reality with which the social

anthropologist is concerned … is not any sort

of entity but a process, the process of social

life.’ The analogy, then, is not between

society and organism as entities, but between

social life and organic life understood as

processes. It was precisely this idea of the

social as a life-process, rather than the idea of

society as an entity, that Radcliffe-Brown

drew from the comparison. And it was for

this reason, too, that he compared social life

to the functioning of an organism and not to

that of a machine, for the difference between

them is that the first is a life-process whereas

the second is not. In life, form is continually

emergent rather than specified from the

outset, and nothing is ever quite the same

from one moment to the next. To support his

processual view of reality, Radcliffe-Brown

appealed to the celebrated image of the Greek

philosopher Heraclitus, of a world where all is

in motion and nothing fixed, and in which it

is no more possible to regain a passing

moment than it is to step twice into the same

waters of a flowing river.

What his critics could never grasp was that in

its emphasis on continuity through change,

Radcliffe-Brown’s understanding of social

reality was thoroughly historical. Thus we

find Evans-Pritchard, in his 1961 Manchester

lecture, pointing an accusing finger at

Radcliffe-Brown while warning of the dangers

of drawing analogies from biological science

and of assuming that there are entities,

analogous to organisms, that might be

labelled ‘societies’. One may be able to

understand the physiology of an organism

without regard to its history – after all,

horses remain horses and do not change

into elephants – but social systems can and

do undergo wholesale structural transform-

ations. Yet a quarter of a century previously,

Radcliffe-Brown had made precisely this

point, albeit with a different pair of animals.

‘A pig does not become a hippopotamus …

On the other hand a society can and does

change its structural type without any breach

of continuity.’ This observation did not

escape the attention of Lévi-Strauss who, in a

paper presented to the Wenner-Gren

Symposium on Anthropology in 1952,

deplored Radcliffe-Brown’s ‘reluctance

towards the isolation of social structures

conceived as self-sufficient wholes’ and his

commitment to ‘a philosophy of continuity,

not of discontinuity’. For Lévi-Strauss had

nothing but contempt for the idea of history

as continuous change. Instead, he proposed

an immense classification of societies, each

conceived as a discrete, self-contained entity

defined by a specific permutation and

combination of constituent elements, and

arrayed on the abstract coordinates of space

and time. The irony is that it was from Lévi-

Strauss, and not from Radcliffe-Brown, that

Leach claimed to have derived his model for

how anthropological generalisation should

be done. Whereas Lévi-Strauss was elevated as

a mathematician among the social scientists,

the efforts of Radcliffe-Brown were dismissed

as nothing better than ‘butterfly collecting’.

Yet Lévi-Strauss’s plan for drawing up an

inventory of all human societies, past and

present, with a view to establishing their

complementarities and differences, is surely

the closest thing to butterfly collecting ever

encountered in the annals of anthropology.

Unsurprisingly, the plan came to nothing.

The full text of this lecture will be published inProceedings of the British Academy, volume 154.

Tim Ingold is Professor of Social Anthropologyat the University of Aberdeen. He is currentlywriting and teaching on issues at the interfacebetween anthropology, archaeology, art andarchitecture.

Claude Lévi-Strauss, Corresponding Fellow of theBritish Academy